A new generation of collaborative software that allows users to retain ownership of their data.

Since the disastrous launch of the RTX 50 series, NVIDIA has been unable to escape negative headlines: scalper bots are snatching GPUs away from consumers before official sales even begin, power connectors continue to melt, with no fix in sight, marketing is becoming increasingly deceptive, GPUs are missing processing units when they leave the factory, and the drivers, for which NVIDIA has always been praised, are currently falling apart. And to top it all off, NVIDIA is becoming increasingly insistent that media push a certain narrative when reporting on their hardware.

The One Big Beautiful Bill Act revives bonus depreciation full expensing, reshapes R&D tax credits, 174 expensing, cost segregation, and energy incentives under 179D and 45L.

https://flak.tedunangst.com/post/can-an-email-go-500-miles-in-2025

Everything around LLMs is still magical and wishful thinking

As a supporter, I would love not to feel this way

https://www.filfre.net/2025/07/everquest/

Meteosat Third Generation Sounder 1 is successfully launched.

TOKYO -- Research papers from 14 academic institutions in eight countries -- including Japan, South Korea and China -- contained hidden prompts direct

https://www.marketwatch.com/story/sleeping-beauty-bitcoin-wallets-wake-up-after-14-years-to-the-tune-of-2-billion-79f1f11f

The title of this post is somewhat misleading. Local-first software rarely needs to be scaled at all.

Apple has just released an AI model that, rather than generating code from left to right, does it out of order and all at once. Here's how.

https://initcoder.com/posts/37-year-old-learning-cs/

tl;dr: If you're an EU citizen, sign the Stop Killing Games initiative here. Or, if you're in the UK, sign this petition.

A month ago, I had a second video on self-hosting taken down. YouTube said it was 'harmful or dangerous content'. I appealed that, but my appeal was rejected. Luckily, I have enough reach I was able to get a second chance. Most content creators aren't so lucky, and content about tools like Jellyfin and LibreELEC get buried over time.

Documenting the evolution of macOS system icons over the past several decades.

: Crims have cottoned on to a new way to lead you astray

https://www.reddit.com/r/AO3/s/67nQid89MW

https://serd.es/2025/07/04/Switch-project-pt3.html

Drilling beneath the hype to expose the truth

Elon Musk is obsessive about the design of his supercars, right down to the disappearing door handles. But a series of shocking incidents – from drivers trapped in burning vehicles to dramatic stops on the highway – have led to questions about the safety of the brand. Why won’t Tesla give any answers?

A new study reveals that air pollution, traditional herbal medicines and other environmental exposures are linked to genetic mutations that may contribute to the development of lung cancer in people with no or hardly any history of smoking.

https://twitter.com/WatcherGuru/status/1941167512491864554

on starting, doing, being, and becoming.

凯撒沙拉是一种风靡全球的经典菜肴，其起源可追溯至1924年的墨西哥边境城市蒂华纳。意大利移民凯撒·卡迪尼在此开设餐厅，一次因食材短缺而即兴创作了这款沙拉，并在客人面前进行桌边服务，意外获得成功。尽管其真正的发明者存在争议（有说法归功于其兄弟亚历克斯·卡迪尼或餐厅厨师），但历史学家一致认为它诞生于蒂华纳。

核心配方与制作 传统凯撒沙拉的核心是整叶罗马生菜、蒜香面包丁以及由生蛋黄、凤尾鱼（或伍斯特酱）、大蒜、第戎芥末、柠檬汁、盐、胡椒和橄榄油制成的浓郁酱汁，最后刨上帕玛森芝士。传统做法注重桌边服务的戏剧性体验，厨师会在木碗中依次加入配料并现场制作。现代多数版本采用亚历克斯·卡迪尼的改良配方，以凤尾鱼酱取代伍斯特酱，并用柠檬代替酸橙。

现代演变与品尝推荐 凯撒沙拉因其基础且灵活的配方，衍生出众多变体，例如洛杉矶Bar Etoile将其与鞑靼牛肉结合。以下是一些著名的品尝地点：

• 凯撒餐厅（蒂华纳）：沙拉的诞生地，仍提供标志性的桌边服务体验。
• 天堂角度假村（圣地亚哥）：提供所谓的“原版凯撒”，用烤蒜和酸橙汁改良，并可添加鸡肉、三文鱼或虾等蛋白质。
• Lola 55（圣地亚哥）：提供带有帕斯拉辣椒盐风味、具有墨西哥特色的凯撒沙拉。
• 丹·塔纳餐厅（洛杉矶）：自1964年开业，是少数仍提供原版桌边凯撒沙拉体验的餐厅之一。

running on jurassic park's fumes for over 30 years

https://sarahrenaeclark.com/lightfast-testing-pencils/

Not just #[derive(Clone)], but all of the standard trait derives are

The Miyawaki Method of micro-forestry is a viral sensation: sprouting tiny, dense, native tree cover in neighbourhoods all around the world. With the promise of afforestation at a revolutionary speed, this planting technique has become the darling of green-space enthusiasts, industry, and governments alike —yet few professional or academic ecologists have commented on its efficacy, or even seem to have heard of it! In this episode, we debate the legacy of Dr. Akira Miyawaki: the man, the myth, and the method.

PHP协程探索

协程是一种可以暂停和恢复执行的函数，能保持内部状态并支持双向值传递。PHP通过**生成器（Generators）和纤程（Fibers）**支持协程。

协程核心概念

• 暂停与恢复：协程可主动暂停执行，将控制权交还调用者，需外部显式恢复。
• 返回与接收值：暂停时可返回值，恢复时可接收值，实现双向通信。
• 状态保持：协程内部变量在暂停期间保留，恢复时可继续使用。
• 类型：分为非对称（只能返回调用者）和对称（可传递控制给其他协程），以及无栈（仅顶层函数可暂停）和有栈（嵌套函数可暂停）。

PHP中的实现

1. 生成器（Generator）

• **PHP 5.5+**引入，通过yield关键字实现协程。
• 非对称无栈协程：只能返回控制给调用者，无法从嵌套函数暂停。
• 双向通信：current()获取返回值，next()恢复执行，send()发送值给协程。
• 主要用途：作为惰性迭代器，节省内存，适用于大数据处理和CLI工具。

2. 纤程（Fiber）

• **PHP 8.1+**引入，基于Fiber类。
• 非对称有栈协程：可从嵌套函数暂停，支持更灵活的代码结构。
• 控制方法：start()启动，resume()恢复，Fiber::suspend()暂停，throw()抛入异常。
• 适用场景：协作式多任务（如并发处理）、状态机、复杂异步工作流。

生成器与纤程对比

协程应用场景

1. 状态处理：如命令行交互工具，通过暂停保持用户输入状态。
2. 惰性迭代：生成器逐项生成数据，避免一次性加载大量数据到内存。
3. 协作式多任务：纤程实现非阻塞并发，如交替执行烹饪和清洁任务，提升资源利用效率。需配合事件循环（如Revolt、ReactPHP）管理协程调度。

总结：协程通过暂停/恢复机制优化流程控制。PHP中生成器适合迭代和内存优化，纤程适用于需嵌套暂停的复杂异步场景。实际应用涵盖CLI工具、数据处理流水线和并发任务。

Gremllm 概述

Gremllm 是一个基于大语言模型（LLM）的 Python 工具类。它的核心思想是：对象的每个方法调用和属性访问都会通过 LLM 实时推理来决定并生成要执行的代码，从而赋予对象动态、即兴的行为。

安装

通过 pip install gremllm 安装。

核心用法

创建 Gremllm 实例时，需要传入一个描述对象性质的字符串（例如 'counter'）。随后，可以对实例调用任何方法或访问任何属性，即使这些方法在代码中并未预先定义，LLM 会根据上下文推理并生成相应的代码。

from gremllm import Gremllm
counter = Gremllm('counter')
counter.value = 5
counter.increment() # LLM 将推理并执行“递增”的代码

主要特性

• 动态行为：通过 LLM 推理在运行时动态实现方法和属性。
• 湿模式：方法调用返回的是活生生的 Gremllm 对象（而非普通值），支持无限链式调用。
• 详细模式：启用 verbose=True 可查看 LLM 为每次调用生成的具体代码。
• 多模型支持：支持 OpenAI、Claude、Gemini 或通过 llm 库使用本地模型。
• 继承：子对象自动继承父对象的 wet 和 verbose 设置。
• 智能错误处理：在库不可用或代码执行失败时具备优雅的回退机制。

配置 LLM

可以通过 llm 库设置首选 LLM 的密钥，或在创建实例时通过 model 参数指定模型（例如 gpt-4o、claude-3-5-sonnet-20241022、llama2）。

示例说明

• 计数器示例：演示了如何创建一个计数器对象，动态调用 increment()、add_seventeen() 等方法，并访问动态生成的 value、current_value 等属性。
• 购物车示例：演示了一个购物车对象，动态执行 add_item、calculate_total、clear 等操作。
• 湿模式示例：在湿模式下，调用 increment() 会返回一个代表数字的 Gremllm 对象，从而可以继续对其调用 double()、square() 等方法，实现无限链式交互。

其他

项目文档指出该工具“居然真的能用”，并推荐阅读一篇关于代码中“Gremlins”概念的背景文章。

We’ve discussed SIMD and vectorization extensively on this blog, and it was only a matter of time before SIMD (or vector) functions came up. In this post, we explore what SIMD functions are, when they are useful, and how to declare and use them effectively. A SIMD function is a function that processes more than… Read

https://www.science.org/doi/10.1126/sciadv.adt3813

prompt engineering is nothing but an attempt to reverse-engineer a non-deterministic black box for which any of the parameters below are unknown:

• training set
• weights
• constraints on the model
• layers between you and the model that transform both your input and the model's output that can change at any time
• availability of compute for your specific query

🎹 Memorize piano scales with ease! A piano practice program w/ MIDI support. Consider it an interactive reference manual - ZaneH/piano-trainer

Berry is a ultra-lightweight dynamically typed embedded scripting language. It is designed for lower-performance embedded devices.